A few weeks ago I was lucky enough to have a friend direct me to the saucy scandal surrounding Green Pasture’s fermented cod liver oil. Curious to see how a bottle of liquid fish organs could set the internet awash in a frothy chatter, I anxiously read through Kaayla Daniels’ report called “Hook, Line and Stinker: The Truth About Fermented Cod Liver Oil.”
It was like warm pop-tarts on a frigid January snow-day.
I work in the natural products industry and spend a lot of my time helping companies sort out their supply chain issues, so the chance to breakdown NMR graphs and cross reference PV values is like waving a bacon-crusted pork chop in front of a pit bull with the munchies. An afternoon well spent.
However, when I was finished my feelings were bifurcated: I loved the dense lab reports, but I felt like I was reading something birthed into existence by the invisible hand of Mike Adams.
No doubt the report contained a lot of truth, but I couldn’t shake my tingling spider-sense that there was a bit of truth-i-ness layered in for dramatic effect.
When it comes to their ability to be manipulated by the media, lab reports and scientific studies sleep in the same bed as valuable sources of raw data that can easily be pornographied to overrepresent the validity of subtle truths.
As advocates of a high-fat diet, I’m sure you can all shake your fists at the head-slapping moments of frustration you’ve had explaining to your vegetarian friends that animal products are not the sole cause of heart disease, the obesity epidemic, type II diabetes, the rising divorce rate, Brian Adams, holes in the ozone layer, Tommy Wiseau’s cinematic tastes, or the John Stewart Show finally going off the air.
The outcomes have causal density, making it difficult to decipher what causes what.
So with that said, it might behoove us to take a closer look at the report to determine where it’s valuable, where it’s ambiguous, and where it might lead us down false trails.
A Word To The Wise
This article is not meant to be a defense or prosecution of any of the parties involved.
Up until I read the report I had little familiarity with the products, people or companies mentioned in the report (with the exception of the labs). I wouldn’t recognize David Wetzel if he walked up to me and picked my nose.
It’s also not meant to be the “definitive-guide-to-every-single-little-thing-in-Kaayla-Daniels-report” because that’s beyond my paygrade. I know just enough about fermentation to be dangerous, and I’d need a 24 Hour Energy to educate myself on the un-abbreviated history of Weston Price’s opinion on butter.
‘Activator X’ sounds to me like it ought to be a brand of cologne, not a pseudonym for vitamin K. So sections VI and VII are getting skipped as well.
Rather, what I want to do is highlight the important points brought up in the document and provide additional context to allow you to make your own decisions about what they may or may not mean.
So with that said, what are the relevant details we ought to file in the back of our heads while perusing this explosive document?
WELL LET”S FIND OUT!
Based on the information contained within Ms. Daniels report can we safely conclude the oil was rancid?
Let’s begin with the fact that rancidity is a term that has a colloquial definition (something that smells icky and initiates a gag reflex) distinctly separate from its scientific one (oxidation) and using the presence of one to infer the existence of the other is a non-sequitur.
Let’s continue with an acknowledgement that the data from the report on the oil’s rancidity is a mixed bag.
Here’s the abbreviated version of what the test results found:
PV: Peroxide value, very low.
AV: Anisidine value, very low.
TBA: Thiobarbituric acid test, very low.
TBARS: A more involved version of the TBA test, came back very high at 23.6.
Fatty Acid Value: The amount of free fatty acids in the mixture. These came back very high, between 14% and 40%.
There was also a TOTOX and Acid Value, but these figures are derived from the tests above, so it suits us to disregard these. (Ie, if the PV and AV value are low then the TOTOX value has to be low as well, giving us no reason to analyze it).
Fatty acid oxidation happens in two stages, primary and secondary. Primary oxidation happens when the double bonds in the fatty acids are oxidized into hydroperoxides. Secondary oxidation occurs when the hydroperoxides from primary oxidation are oxidized into various ketone bodies, epoxides and other metabolites of the primary phase..
Because the stages of oxidation happen at different times it’s possible you can have measures for one look fine and still have rancidity because the other phase of oxidation has taken its place.
PV and to a lesser extent fatty acid value are measures of primary oxidation. AV, TBA and TBARS are measures of secondary oxidation.
The curious issue is how you should interpret the body of evidence given that some of the values in each category were high while others were very low.
TBARS stands for the Thiobarbituric Acid Reactive Species test. When thiobarbituric acid reacts with something it turns pink and when the test is finished you measure how much UV absorption occurs within 532-535 nm1, which ought to measure the amount of TBA reactants in the mixture.
The test isn’t very accurate because TBA reacts with a lot more than just byproducts of secondary oxidation and tends to produce false positives. With standard fish oils it has a standard deviation of about 15%2, and with a complex mixture you’d expect this number to be higher.
All by itself TBARS is not considered a good way to judge rancidity or much of anything else due to its inaccuracy, even though it does provide some useful information.
Ideally you’d want to diminish the error in a TBARS test by running the sample through a procedure called HPLC, which is a process that’s useful for separating molecules in a complex mixture. The separation allows you to tell the difference between products of the reaction that truly represent oxidation and those that are just oily-based flotsam & jetsam.
And even then…….the test still runs amok. The use of the two methods together has been explored and even with the addition of HPLC it isn’t possible to completely eliminate the possibility of unwanted TBA reactants.3
Ms. Daniels herself alludes to the unreliability of the TBARS test on page 22 of her report:
“Given that FCLO naturally contains some protein residue from “fermenting” livers, TBA would appear to be an unreliable way to test FCLO for rancidity.”
And she’s basically right. TBARS just jumps all over the place, even when you try and take measures to calm it down.
Free Fatty Acids
This leaves the issue of the free fatty acid values, which are unarguably high. The figure indicates what percentage of the fatty acids have been hydrolyzed from the triacylglycerol backbone, with the logic being that a free fatty acid is blazin’ down the rancidity superhighway on its way to Beelzebub’s in the seventh circle of fatty acid hell.
However, if we’re careful, we’ll see there’s an important assumption being made by stating rancidity and the presence of free fatty acids are doppelgangers of one another: a fatty acid is hydrolyzed en route to being oxidized, therefore the presence of a free fatty acid is proof of oxidation.
If we harken back to our university philosophy courses we should recall this is an example of the Roostery Syndrome: assuming the coincidence of two events implies one caused the other.
The question then, is how do we interpret an oversized risk indicator of oxidation when the other evidence does not paint a daunting picture that it took place?
One interpretation, the one taken by Ms. Daniels, is the scorched-earth scenario where the byproducts of primary and secondary oxidation have withered away to nothing, leaving only a cesspool of free fatty acids in their place.
On page 23 of her report she explains her interpretation of the Green Pastures Acid Value as thus:
“these Acid Value numbers are extremely high. So high that they blow any claims that FCLO is a non-rancid oil right out of the water.”
That’s one opinion you can take, and it’s not illegitimate.
However, there are additional points we want to consider before we decide to use the rest of our fermented cod liver oil as anti-freeze in our car engines:
● EPA/DHA Value. The EPA and DHA content in the oil were appropriate, and it’s unlikely that you’d have highly oxidized free fatty acids and high levels of EPA/DHA at the same time, which are themselves extremely sensitive to oxidation.
● THE PV was fine. Twice. The hydrolysis of free fatty acids from the glycerol backbone usually coincides with primary oxidation, and the extremely low PV values create the possibility that the FFA value is a red herring. This point is strengthened by the fact that the NMR data on lab report #5 (pg. 91-92) returned oxidation values that were low to the point of being undetectable. This is a useful observation because an NMR test is non-invasive and wouldn’t have the same errors that a traditional PV test would have, which measures the amount of material that reacts with potassium iodide. (This point about the NMR data was also not brought up anywhere in Ms. Daniels’ report).
● Acids….run amok! The method used to determine the free fatty acid value is an acid-base titration4, and the presence of acidic substances besides fatty acids would cause the numbers on the test to err on the high side. I have no idea how Green Pasture ferments their livers, but the majority of fermented products use some sort of starter material to initiate the process that contains acidic material. That, and acidic byproducts released during the fermentation process would cause the FFA value to falsely skip up a percentage or four.
Of course without knowing exactly how Green Pasture ferments their products it’s impossible to say how much or how little the last point distorts the free fatty acid value. But taken together these observations ought to at least introduce the possibility that a high fatty acid value doesn’t indicate much more than the complexity of the process used to create the oil.
I wish I possessed a pistol-whippin’, whizz-bang insight that’d totally reverse the low vitamin content reported in lab reports 1, 2, 4, 6 and 8.
Sadly, I cannot. At least nothing that wasn’t brought up in the excellent articles by Sally Fallon and Chris Masterjohn. So to avoid being redundant I’ll recommend you read them if you want to know their opinion on the issue. (I’ll forgive you for clicking on the linky, so long as you promise to come right back).
However, there’s an under-discussed point about the fragilities of the tests used to measure their concentration, which I’ll bring to your attention here.
On page 36 of her report Ms. Daniels states:
“If the method is UV detection, it is very easy to be fooled.”
She was using this as an indictment of Green Pasture’s testing methods, but as a rhetorical point it’s ironic because UV detection was the method used to determine the vitamin A content in lab reports 2 and 8.
The method can be found used was a variant of UV-Vis spectroscopy called colorimetry, which uses a colorimeter to measure the concentration of a solution by its absorption of a range of light. 5
In general UV detection is best used for analyzing raw materials. When you’re dealing with finished substances, particularly a complex one, it’s very easy for the presence of other compounds to throw off the amount of light absorbed and create inaccurate results.
To be fair, it tends to skew results high and not low so it’s possible the vitamin A content is actually lower than what’s stated, but the method has a high level of variance overall and isn’t well suited to final product testing.
To detect the levels of vitamin D used in the mixture used in lab report #1 the method employed was LC/MS, which is a combination of HPLC (high pressure liquid chromatography) and Mass Spectrometry.
LC/MS is to analytical chemistry what snapchat is to social media: the hip new toy that all the fanboys flock to, even when you’re pretty sure Facebook works just fine.
HPLC is the method most commonly used to detect final product concentrations and it’s usually accurate unless sample sizes are very small. (For example HPLC can usually detect vitamin D down to about 0.5 iu/g, and below that you’d need to use LC/MS).
The appeal of LC/MS is that by combining the two methods it helps erase the faults created by either one on their own. Mass spectrometry can’t separate stereoisomers (compounds that have the same molecular weight but different geometries) and HPLC doesn’t always provide useful information about what a chemical actually is after it’s been separated.
So by separating chemicals through HPLC and then passing them through a mass spectrometer you get the best of both worlds: highly sensitive separation and useful data about its molecular weight.
And to be honest…….yes, it’s cool.
However, LC/MS has a tendency to undershoot concentrations due to problems with the handoff from the column to the mass detector.
Mass detectors can only detect molecules that are electrically charged (ie, ions) and they have to pass through the detector in a complete vacuum, so all the solvent from HPLC needs to be removed and the mixture needs to be entirely ionized to ensure clean results.
If this isn’t handled properly the results will produce false negatives: an absence of substance that’s actually there.
Of course, these complications will skew results by an order of approximation, not an order of magnitude. There’s small chance you could apply an equally legitimate test of concentration and come up with something radically different, unless the lab techs doing the mass spec had performed radical IQ-lowering practices beforehand………like sitting through a 24 hour marathon of The Room.
I think if there is significantly more vitamin D activity than what’s being reported it’s due to the presence of various vitamin D metabolites that are not showing up in tests for cholecalciferol, not an inaccurate measure of cholecalciferol or ergocalciferol itself.
Variation: It’s Natural
To be honest I have little familiarity with the marketing claims made by Green Pasture on their website or in their literature about their vitamin content. When I go there to look at their product labels I don’t see any mention of vitamin content so I’m not sure what caused the impression the vitamin levels would be higher than they were.
However, now’s a good time to point out an inconvenient truth about putting unprocessed products into a bottle, storing them in warehouses for months at a time and then shipping them across the country: mother nature doesn’t like to be standardized.
If you’re hell bent on getting a consistent nutrient content from a truly natural product with 99.84738495847% accuracy you need to waive a big “Goodbye!” to green pastures and waive a big “Hello!” to white lab coats.
Lickity-split vitamin levels are only possible if they’re assembled into something synthetically. Ocean waters and grasslands will never do the trick.
I work a lot with botanicals, and the pharmacologically active plants will often have their concentration of active compounds vary throughout the day depending on the weather.
A good example would be St. John’s Wort. Its active ingredients are hyperforin and hypericin and both are very photosensitive. (This might be one reason it increases photosensitivity among people who take it). Hyperforin is oxidized by light and will easily be destroyed if exposed to too much UV radiation. Hypericin is activated by light and its biological activity depends on a double proton-transfer that takes place after UV exposure to be physiologically useful.
If you want to optimize your crop for hypericin you need to harvest it during the afternoon. If you want to optimize for hyperforin you ought to do it early morning or early evening. And if you’re not careful with the plant immediately after it’s taken out of the ground you might as well have grown it on Mars because it’ll be degraded beyond recognition.
This dynamic isn’t unique to St. John’s Wort either. Ashwagandha, Shilajit, Gingko Biloba, Rhodiola and most other biologically active herbs have active compounds that don’t take well to the vagaries of Mother Nature.
Granted, I know less about fish than I do herbs, but assuming fat soluble vitamins are stored and utilized at different rates according to diet, season and environmental conditions I see no reason you wouldn’t have large swings in vitamin content from batch-to-batch.
The identity of the fish species used in the oil was called into doubt for the following reasons:
● A DNA test identified it as Alaskan pollock.
● The oil had a high amount of EPA relative to DHA, which is unusual for Cod.
● The oil had a high amount of trans fatty acids, also unusual for Cod.
Most people probably think of the DNA test as the most damning, but I consider it the piece of evidence most easily thrown away.
The amount of DNA needed to get a reliable signal is below the amount typically found in oils,6 rendering the test moot.
It’s known colloquially as ‘DNA barcoding’ and it involves comparing DNA sequences for a particular gene with a database of known variations for different species. It’s done with mitochondrial DNA because it has a higher turnover rate than nuclear DNA, allowing for more differentiation among species. The most popular gene to study is the one that codes for the protein cytochrome oxidase 1, and this is what was sequenced in lab report #3.
Ms. Daniels herself acknowledges the unsuitability of DNA tests for the product in question on page 48 of her report (emphasis added by me):
“DNA technology can now accurately distinguish all common fish and seafood as well as most exotic species. Additionally, it can identify species of fresh, canned and frozen fish in readymade products so long as they are present at levels greater than ten percent. Unfortunately, DNA procedures do not work well with oils.”
So instead, according to her report she used remnants from the cattle lick product sold by Green Pasture to get the test done.
It’s up to you how much relevance you want to place on a DNA test that used a sample that wasn’t from the product being questioned. Not from a different batch mind you, but from something that’s categorically different and not branded as cod liver oil.
I don’t have David Wetzel’s number on speed dial so I can’t speak to the relationship between the livers in his cattle lick product and the oil in his FCLO, but nonetheless I think this registers as a pretty big caveat about how we ought to interpret the permanence of the results in lab test #3. The test wasn’t done on the oil!
In the absence of a good DNA sample NMR is the best way to test for species identification. The procedure measures how different molecules in a nuclei jump around in response to electromagnetic radiation which provides useful information about a compound’s purity and structure.
In particular, you can measure the specificity of various fatty acids at the sn-2 location of the triacylglycerol molecule, which is a reliable species indicator since it can’t be gamed and is distinct from fish to fish.
The results from lab report #5 found the DHA specificity to be 82% and 44% for EPA, which corresponds to what you’d expect to find.
The only abnormality in the result was an unusually large amount of free fatty acids, which shouldn’t be a surprise given the other lab reports.
The results of the test were summed up like this:
“These values of sn-2 position specificity are similar to previous analysed cod liver oil samples…….the overall 13C NMR carbonyl profile are similar to cod liver oil, although the levels of monounsaturated fatty acids in sn-2 position seems to be a bit higher than previously analysed cod liver oils……..The sample is more similar to cod liver oil of wild origin than e.g. previous analysed salmon oil or anchovy oil“
Ie, with the exception of a higher than normal FFA content, the results were basically fine. Or at least, not enough to assume fraud.
Fatty Acid Ratios
The hardest result to explain away is the EPA:DHA ratio found in the fish and its high trans fat %.
It was about 2:1, which is very unusual. It’s also difficult to interpret the results since there’s not a lot of codified information about what the values ought to be.
The closest thing there is to a reference on the fatty acid levels in various fish is the book “Long Chain Omega-3 Specialty Oils”, and with a sticker price of $220 it’s not exactly priced to be an a best seller. (And no, I didn’t buy it).
Most of the feedback I got for this section was from a conversation with a QC analyst at GOED, which is a non-profit organization that monitors the fish oil industry. (He preferred to remain anonymous).
After explaining the issue to him this is what he said:
● Anchovies are the only wild fish that regularly has more EPA than DHA
● Tests for fatty acid content usually underestimate DHA content and it’s usually best to have them performed multiple times to get a good reading, because they all undershoot.
● Alaskan pollock has an EPA:DHA ratio that’s closer to what was observed in the lab report
● The oil was unlikely to be anchovy oil because it has an EPA ratio that’s closer to 17%-18%, rather than the 13.5% observed in the sample.
● It’s possible the fish were fed anchovies at some point, although this would not coincide with the NMR results which suggested the oil was from wild cod.
● Trans fat percentages can sometimes measure oxidation products moreso than trans fats themselves
When you add all this up, I think you get a picture that’s capital-A Ambiguous. An unreliable DNA test on a non-product sample, an NMR test which more or less came out fine, and a fatty acid test that contradicts the results of the other two.
In the best-case scenario the reported DHA levels are artificially low due to testing conditions, which makes everything hunky-dory. It’s possible the oil really is from pollock (which is not an uncommon occurrence), or even weirder……there’s some other hidden variable in the supply chain that hasn’t been uncovered yet.
There’s a side debate about the relevance of something being classified as pollock or cod, since they both belong to the same taxonomic family, but that’s a fish to fry for another day (PUN intended).
Friends, if you’re lucky the joy you got from reading this was at least ⅕ the entertainment value I got from writing it.
You know what they say……..explosive exposés of esoteric health products only come around once……so when your number’s called, ya gotta’ be ready to go.
I am not the Arbiter of Truth here, so please don’t mistake the words or assumptions in this article as rapture.
In fact, I’d hope the take away message is that testing is fragile. The method used is only as reliable as the person implementing it, and it’s entirely normal for different labs with different equipment using different methods to come up with very different results even if they’re trying to measure the same thing.
It’s the nature of the beast. And with many of the more exotic methods, you’re charging into territory that has not yet been formalized, making it hard to draw firm conclusions. Of course that’s what makes it interesting, but I loathe to think that some lab reports on fish oil can provide tinder for our tribalist animal spirits to run rampant.
But of course on the internet, anything’s possible.
Have a nice day!
1). Kishida, et. al. “Re-evaluation of malondialdehyde and thiobarbituric acid-reactive substances as indexes of autoxidation based on oxygen consumption”. http://pubs.acs.org/doi/abs/10.1021/jf00025a001
2). Samb, Norveel Thea. “Analytical Methods for Determination of the Oxidative Status in Oils”. http://www.diva-portal.org/smash/get/diva2:536470/FULLTEXT01.pdf
3). Chirico, et. al. “Lipid peroxidation in hyperlipidaemic patients. A study of plasma using an HPLC-based thiobarbituric acid test.” http://www.ncbi.nlm.nih.gov/pubmed/8225034
4). AOCS Official Method Ca 5a-40
5). AOAC 974.29
6). Martinez, I et. al. “Destructive and non-destructive analytical techniques for authentication and composition analyses of foodstuffs” http://www.sciencedirect.com/science/article/pii/S092422440300164X
Byline: Jonathan Bechtel is the owner of Health Kismet, a dietary supplement manufacturing and consulting company. He sources, develops and markets natural products both for himself and others.